Greg,
I agree that the first issue we face is whether 822-PEM should
become syntactically compatible with MIME-PEM, so that interoperation
without gateways will be possible. I would not say that the 822-PEM
should be supplanted by MIME-PEM since it is reasonable to have a
subset defined for 822-PEM which does not require even a basic MIME
mail reader. But a level of syntactic commonality (which is what has
been proposed previously but not throughly worked out) is a likely
compromise.
I agree with your observation that the MIME-PEM effort was not
constrained to produce solutions which are backward compatible with
822-PEM. Nonetheless, a design which minimizes changes to the
relatively small number of existing domestic and foreign PEM
implementations is a possible goal for the WG to consider. No
question that I get lots more MIME mail than PEM mail, but that's
largely because the ID and RFC announcements come that way. The MIME
mail I receive isn't doing anything special for me (especially since I
don't have a MIME mail reader!). Nonetheless, I agree that a
strategy that allows 822-PEM to be syntactically compatible with
simple PEM-MIME strikes me a desirable approach.
For example, to facilitate this, one might be a bit more
spevcific than the current proposal in the choice of
Content-Transfer-Encodings for MIC-ONLY messages. Simple syntactic
substitutions (e.g., MIME boundaries for 822-PEM boundaries) may be an
easy change so long as everyone is convinced of the equivalence
between the two. The bodypart ordering discussion seems to be one of
implementation ease vs. perceived user annoyance, and will certainly
be the subject of some (value laden) discussion next week.
The Content-Domain vs. Content-Type issue may revolve around
the idea of a basic PEM capability which can be profiled for not only
MIME but other messaging environments. For example, if one transports
PEM traffic in an X.400 bodypart, it is important for PEM to carry the
identification of the its content. The WG can consider adopting MIME
syntax for this, but everyone needs to be comfortable that the
resulting PEM capability has not lost any functionality outside of
MIME environments. I'd like to think ahead to MIME-X.400
interoperability as well as back to MIME-822.
Well, thanks for your quick response to my analysis. It was
offered as a starting point for the discussion next week and I hope
you will be able to attend and participate in the discussions.
Steve